INTRODUCTION
Total knee arthroplasty is a high volume, high quality procedure with nearly 900,000 procedures performed annually in the United States and 3,416,000 projected by 2040 (Singh et al. 2019). While complication rates remain low and improvement in patient quality of life years (QALYs) remains high, there is still significant room for improvement (Konopka et al. 2018). Given the high volume of procedures, revision TKA is growing rapidly with over 70,000 reported annually (Upfill-Brown et al. 2022). While patient outcomes are considered good, some studies still report up to 20% of patients are not satisfied with their TKA (Bourne et al. 2010; Noble et al. 2006). With ever improving implants and techniques, the indications for TKA are expanding, leading to a younger, more active population placing further demand and expectations on TKA. With the increase in patient demand and improved implant durability, we are still measuring success and failure in the same antiquated way, using knee range of motion (ROM), physiotherapy (PT) reports, and patient-reported outcome measures (PROMs), which are increasingly being seen as problematic and potentially not accurate in reporting patient outcomes (Halawi et al. 2020).
One objective tool that can measure the patient recovery during the TKA process is Persona IQ, (Zimmer Biomet, Warsaw IN) with the CANARY canturio™ tibial extension (Canary Medical, Vancouver Canada). This smart implant system, which received FDA market release authorization on August 30, 2021, contains an inertial measurement unit (IMU) with a gyroscope and accelerometer. The system utilizes the Canary Health Implanted Reporting Process (CHIRP®) System (Canary Medical, Vancouver, British Columbia), connecting the implant via the cloud to the patient and surgeon. The embedded IMU allows step count, walking speed, distance, stride length, cadence, functional knee range of motion (knee range of motion while walking), and tibia range of motion to be remotely collected, analyzed, and displayed in the form of gait parameters via cloud connectivity to the patient and heath care team.
The availability of gait parameter data from a large cohort of patients already implanted with a smart knee allow a patient’s daily gait parameters to be plotted as a percentile score, affording the opportunity to see their individual progress in the context of their demographically match TKA peers. The goal of remotely monitoring patients’ recovery of gait and activity is to improve efficiency in care, increase patient and physician engagement, and, potentially, prevent impending complications, before they arise or get worse. Clinicians currently lack a consistent and defined, evidence-based algorithm for assessing individual-level recovery after the arthroplasty.
In this case report, we describe a primary TKA patient who was recovering well initially and then declined during the first few weeks following surgery. We were able to identify an issue with her recovery through the use and interpretation of the gait and activity data from a smart implant and proceed with early manipulation under anesthesia (MUA) leading to improvement and normalization of her recovery. MUA procedure is infrequently performed after primary TKA (0.6% in the American Joint Replacement Registry and 2.6% in the Medicare database) (Brigati et al. 2020; Parkulo et al. 2023). There is no consensus on MUA indication, though failure to achieve flexion ≥90o within 4-8 weeks of TKA is commonly cited (Choi et al. 2015; Bawa et al. 2013). Most studies indicate timing of MUA (i.e., within 3 months of surgery) is important (Parkulo et al. 2023; Kornuijt et al. 2018; Gu et al. 2018; Issa et al. 2014; Cates and Schmidt 2009; Bawa et al. 2013), with pre-surgical knee ROM a commonly-reported prognostic factor (Gu et al. 2018; Choi et al. 2015; Ipach et al. 2011).
CASE PRESENTATION
Patient
This patient is a 47-year-old female with severe degenerative arthritis. She had failed numerous years of physical therapy, injections, and NSAID use, and we indicated her for right TKA based on physical exam and radiographs (Figure 1). She was 5’7" tall, 180 pounds, and her preoperative ROM was 10-degree flexion contracture to 115 degrees of flexion for a 105-degree total arc of motion preoperatively. The cemented TKA surgical procedure went well and without complication with robotic balance data as shown (Figure 2). The surgery was done with patella resurfacing, robotic assist, standard medial parapatellar approach, without a tourniquet. The patient had an IPACK and saphrenous blocks for pain control using 0.25% Marcaine and also underwent intra-articular injection intraoperatively using a Ranawat cocktail. She underwent spinal anesthesia using a chloroprocaine spinal and was discharged home on Tramadol, Percocet, Celebrex and 81 mg Aspirin BID after clearing PT 3 hours post index operation, with a 3x per week PT plan. At her 2-week follow-up visit her ROM was 0 degrees extension and 90 degrees of flexion, pain was controlled and swelling was reasonable, radiographs showed no complications with typical kinematic alignment (Figure 3), she had tapered off narcotics and was progressing in PT and doing well. By 4.5 weeks post surgery, she had increased self-reported pain scores using our post-operative application provided to each patient, and her walking speed, stride length and tibial range of motion percentile scores declined (Figure 4). Consistent with her shorter stride length her cadence was fast (~75th percentile) (Figure 4). Her step count continued to follow the 50th percentile curve, increasing week over week. We called the patient in to the office for evaluation, 3.5 weeks prior to her scheduled routine 8-week follow-up visit. She was increasing her activity and it was causing her pain and swelling. She presented with a stiff knee and decreased ROM. She had a 10-degree flexion contracture and was only able to get to 70 degrees of flexion, significantly worse than at her 2-week visit. In discussion with the patient, we decided to proceed with early MUA. This was done with sedation, a saphrenous block with 0.25% Marcaine and we cycled her knee numerous times to make sure she could obtain full ROM (full extension to 135 degrees flexion). Following the MUA, she was instructed to attend PT 5 days a week for 2 weeks followed by 3 days a week for the following 4 weeks. Following the MUA she did well and by 6 weeks post MUA she had full extension and 125 degrees of flexion, her swelling was greatly improved, and she was walking without a limp.
DISCUSSION
We achieved a well-balanced primary TKA surgery, as evidenced in the robotic data, and good two-week clinical outcomes, as assessed at a routine follow-up visit, yet this patient started showing both a decline in objective gait metrics, detected by a change in the patient recovery trajectory, and a noted change in comparison to the normalized patient cohort, shortly after 4 weeks post-surgery. The change in slope of the patient recovery trajectory prompted a phone call, that prompted an office visit and discussion, which prompted an early intervention. This previously unavailable objective data provides an opportunity to understand real-world activities of daily living more comprehensively and detect concerns between routine follow-up visits. In the case of our patient, kinematic data combined with clinical data provided via the patient app in the early post-op period let us know that she was not progressing and prompted early evaluation and the decision to perform MUA.
Early identification and treatment of potential complications is pivotal to both patient recovery and care cost effectiveness. Identifying patients at risk for functional impairment, before revision surgery is needed, has been hindered by both an absence of objective outcome data and the resource-intensive nature of gait analysis in a clinical setting. In this case report we describe a patient who underwent primary TKA with an implant containing embedded sensor technology within the tibial stem that transmitted high fidelity, daily gait data following surgery. The smart implant has the potential to record in vivo data that can track recovery and provide adjunctive information that assists in clinical decision-making. Unlike traditional in-office performance outcome measures to quantify physical function at discrete timepoints, smart orthopedic devices can provide unprecedented, daily data over long periods (10 years). The smart implant permits remote monitoring of patient gait in their own environment, performing daily activities that are important to them. Data provided by smart sensors are continuous and readily available, making patient monitoring significantly easier and less costly than measurements performed in traditional gait labs.
Alternatives, such as wearable sensors, can be limited by inaccurate alignment of the sensor relative to anatomical landmarks (Cuesta-Vargas, Galán-Mercant, and Williams 2010; Poitras et al. 2019), reported noise, variability and drift in the real-world setting (Matijevich et al. 2020), the potential for loose fixation to the patient, and challenges with consistent patient use so that daily data is collected over weeks. Patients frequently stop using wearables as time from surgery increases (Hermsen et al. 2017; Lyman et al. 2020).
As value-based care initiatives continue to concentrate on patient satisfaction, patient-focused PROMs are becoming common in clinical settings and can provide information on patient functional recovery and satisfaction. Their scientific and diagnostic utility depends largely on what information is captured, how complete the information is, how inclusive they are, and how representative they are of the patient population at large. However, their usefulness may be limited by their subjectivity and reliance on pain rather than movement as a metric of success (Terwee et al. 2006; Stevens-Lapsley, Schenkman, and Dayton 2011). Also, the association has been reported to be weak between PROMs and an existing standard for functional outcomes, standardized tests (e.g., the 400m walk pace, chair stand and 20 m pace tests) (Hill et al. 2023).
The daily spatial-temporal gait metrics transmitted from each patient’s implant allow both patient and physician to chart individual progress throughout recovery for seven gait metrics (including distance). The kinematic data also become meaningful and diagnostic in the context of a large group of existing patients with smart implants. For example, to see if a patient with a smart implant is on schedule with their recovery, the individual’s data for each gait parameter is compared to estimates of the 5th through 95th percentile values within their appropriate cohort (i.e., men <65y or ≥65y; women <65y or ≥65y). The patient can immediately see how they are progressing relative to others in their cohort TKA. In the case of our patient, we could see that, relative to other women less than 65 years old, she failed to progress normally early after surgery and we intervened with an MUA. There is no validated clinical prediction tool to estimate a patient’s recovery trajectory, though post TKA knee flexion guidelines have been reported (Ebert, Munsie, and Joss 2014) and Kittelson et al. published a reference chart for flexion active ROM to assess postoperative recovery for individual patients relative to others who have previously undergone TKA surgery (Kittelson et al. 2020). Implantable, smart sensor technology represents a new era, allowing clinicians to monitor individual patient recovery, alone or in the context of a peer cohort, set appropriate individual recovery goals and, potentially, avoid adverse events.
CONCLUSIONS
Herein we presented a case of an early intervention with MUA following TKA. The need for an MUA was detected early in the post-recovery period, prior to routine follow-up, by smart sensor technology in the implant. Remote monitoring with the smart implant has the potential to improve patient compliance, increase physician and patient engagement, and provide data critical for an early intervention that impacts recovery, as exemplified by this patient.
Acknowledgment
The authors wish to acknowledge and thank Katie Miller, MS, and Max Gill, MBA, for thoughtful review of the manuscript and Jane Bailly, PhD, for editorial assistance.